This project will develop clinically validated multiscale models of cardiac dynamics that integrate fluid dynamics, electromechanical coupling, and fluid-structure interaction (FSI) to simulate intracardiac flows and blood co-agulation in atrial fibrillation (AF). AF is the most common sustained arrhythmia in the U.S. and is associated with serious complications, including thromboembolism and stroke. Anticoagulation is commonly prescribed to patients who have an elevated stroke risk. However, current risk assessment indices, which lack individualization based upon atrial structure or function, classify most AF patients as being at intermediate risk. The core hypothesis of this research is that treatment guidelines using current risk assessment metrics result in many AF patients receiving unneeded anticoagulation and unnecessary monitoring for thrombosis. The long-term objective of this research is to develop new, broad-spectrum approaches to clotting risk assessment in AF that provide personalized risk prediction. The scientific premise of this proposal is that comprehensive models of atrial dysfunction will enable mechanistic studies of flow and clotting in AF that will ultimately facilitate individualized treatment. In AF, most clinically significant thrombi form in the left atrial appendage (LAA). The anatomy of the LAA is extremely heterogeneous, and although there is an emerging appreciation that LAA anatomy affects clotting risk, anatomy is not considered in current guidelines. Computer models provide ideal platforms for studying the impact of structural and functional variations on LAA flow patterns, but most existing cardiac fluid dynamics models focus on the ventricles. Further, no existing FSI model of the atria includes a detailed description of the LAA, which, like the ventricles and unlike the main LA cavity, is highly trabeculated. A key innovation of this project is that it will develop clinically validated FSI models of cardiac flow in patient-specific descriptions of LA anatomies, including realistic models of the LAA. These models will be extended to include biophysically detailed models of coagulation dynamics and clot transport. This project aims both to establish these models and also to apply them to study flows and clotting dynamics in two therapies for AF: (1) percutaneous LAA exclusion via the WATCHMAN device and (2) electrically isolating the LAA in catheter ablation therapy. In the case of LAA exclusion, the incidence of device-associated thrombosis is 3.4%; consequently, post-operative anticoagulation therapy is currently used in all patients receiving these devices. Electrical isolation of the LAA is rarely performed because of concerns about its effect on systolic flow and stroke risk, and the inability to identify patients who would benefit. The core modeling approaches developed in this project can also be deployed to simulate thrombogenesis in a range of significant medical conditions (venous thromboembolism, deep vein thrombosis), medical devices (prosthetic heart valves, ventricular assist devices, IVC filters), and novel biomaterials. Ultimately, models using this platform are expected to be submitted to the FDA Medical Device Development Tools program as non-clinical assessment models to predict pre-clinical device performance in regulatory submissions.